Four clinical research studies will evaluate patients with severe upper gastrointestinal bleeding (UGIB) from peptic ulcers of Dieulafoy's lesions. In studies I-III which are blinded, multicenter randomized trials, our hypothesis is that combination therapy (epinephrine injection +bipolar electrocoagulation-BP)will be significantly better than single therapy (BP) for initial hemostasis, reduction in early rebleeding rates, and reduction in hospital costs of car. Specific aims of Study I are to determine whether; 1) combination therapy stops active bleeding from ulcers or prevents rebleeding from ulcers with non-bleeding visible vessels (NBVV) more often than single therapy, 2) there is any increased risk of combination therapy, 3) there are characteristic profiles of patients with these lesions who are affected most or least by combination therapy, and 4) there are differences in the estimated costs of hospital care (direct and indirect). The specific aims of Study II are 1) to determine whether combination therapy for ulcers with adherent, non-bleeding clots prevents rebleeding more often than medical therapy, 2) there is any increased risk, and 3) there are differences in direct and indirect costs of hospital care. Specific aims for study III are 1) to determine whether combination therapy for Dieulafoy's lesion with active bleeding or non-bleeding visible vessel prevents further bleeding more often than bipolar electrocoagulation alone, 2) there is any increased risk, and 3) there are differences in direct and indirect costs of hospital care. Specific aims of Study IV are to determine 1) the long-term recurrence rates of lesion bleeding (ulcer or Dieulafoy's) and the endoscopic-clinical characteristics of the recurrence, 2) the effect of Helicobacter pylori eradication on rates of ulcer recurrence, ulcer rebleeding and other complications, 3) the effect of non-steroidal anti-inflammatory drug (NSAID) ingestion during follow-up on lesion recurrence and 4) the direct and indirect costs of patients during long-term follow-up. All studies are multicentric and prospective, have data entry onto CURE computer file, and have comprehensive analysis of different outcomes stratified by lesion and treatment. These studies will significantly expand our knowledge about bleeding ulcers and Dieulafoy's lesions; the impact of different endoscopic treatments on routine outcomes and cost of care for these patients; the patterns of recurrence of these lesions; and the long-term response of patients with these lesions to NSAID ingestion and treatment of H. pylori.